基于改进GA的云计算任务调度策略

针对传统遗传算法在云计算任务调度过程中的收敛速度慢和易早熟等问题,提出了一种基于遗传优化算法的双适应度函数改进算法.该算法采用任务完成时间和任务完成成本为双适应度函数,引入个体相似度概念来提高种群质量;采用并列选择法进行选择操作,并且采用自适应规则约束交叉和变异操作,提高种群个体质量,加速进化策略可以有效地避免早熟.结果表明,改进的遗传算法有效地加快了云任务作业调度的收敛速度,并改善了易早熟等现象.

Task scheduling strategy for cloud computing based on improved GA

Focusing on the slow convergence rate and prematurity of traditional genetic algorithm(GA)in cloud computing task scheduling process, an improved double fitness function algorithm based on the genetic optimization algorithm was proposed. The task completion time and task completion cost were used as the double fitness function in the proposed algorithm, and the concept of individual similarity was introduced to improve the population quality. The parallel selection method was used to perform the selection operation, and the self-adaption rules were used to restrain the crossover and mutation operation to improve the individual quality of population. In addition, an accelerated evolution strategy could effectively avoid prematurity. The results show that the improved genetic algorithm can effectively accelerate the convergence speed of cloud task scheduling and alleviate some phenomena, e.g.prematurity.